Evaluation of 2-Week Doctor Visit Reporting in the National Health Interview Survey A new report, "Evaluation of 2-week Doctor Visit Reporting in the National Health Interview Survey" (NHIS) has been released by the National Center for Health Statistics. The report describes the study methodology and presents findings for an evaluation of the reporting of ambulatory medical care visits in the NHIS 2-week reference period, part of a larger evaluation effort called the Health Interview Evaluation Survey (HIES). The HIES is one of many research studies that have examined the accuracy of household survey reports of ambulatory medical visits through record checks. It was conducted by Westat, Inc; the Project Hope Center for Health Affairs shared the design and analysis responsibilities. The study sample was drawn from the membership rolls of the Group Health Association (GHA), whose staff provided essential assistance in identifying the sample and in making available participants' medical records. At the person level, the HIES found that about 78 percent of list-sample persons with one or more GHA visits in the medical record reported at least one visit in the interview, with almost the same proportion (80 percent) of list-sample persons reporting visits having them confirmed by the medical record. For list-sample persons, there was virtually no net differences between the interview and medical record in the number of people with visits. For adult household members not present for the interview, however, less than one-half of the persons with visits in the medical record had visits reported in the interview. Compared with the findings for visits, the findings for telephone calls show a considerable net underreport of GHA telephone calls by all participants. Only about one-third of list-sample persons with calls in the record reported a call in the interview, while about 60 percent of persons with telephone contact reported in the interview had it confirmed by the medical record. The HIES found differences in reporting by respondent characteristics. Younger people were less likely to underreport and more likely to overreport than older people, leading to a net overreport of about 14 percent as compared with the medical record for persons aged 18-44 years, while all other age groups had a net underreport of between 4 and 9 percent. Men were more likely to underreport than women.
This page last reviewed
January 11, 2007
|